nankaicyber
/
OrderVis
135
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update README.md

This commit is contained in:
505025234 2021-11-19 16:08:55 +08:00 committed by GitHub
parent 79360e08d2
commit 58f9eed4f6
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
README.md
View File

@ -1,9 +1,9 @@
# OrderVis # OrderVis
It's a interface for ReorderVis, a tool that can rise the explainability of your Rnn Model. It's a interface of OrderVis, a tool that can rise the explainability of DTM(Deep Temporal Model).
## How to use ## How to use
This system is divided into two parts - browser side and machine learning side. The machine learning side calculates interpretable information of the RNN model the researchers set based on the input and the prediction value of the input reordered by the end user. The machine learning part can run directly to provide part of the interpretability. The code has been placed under the machine learning side branch https://github.com/505025234/OrderVis/tree/machine-learning-side This system is divided into two parts - browser side and machine learning side. The machine learning side calculates interpretable information of DTM that the researchers set based on the input and the prediction value of the input reordered by the end user. The machine learning part can run directly to provide part of the interpretability. The code has been placed under the machine learning side branch https://github.com/505025234/OrderVis/tree/machine-learning-side
While the browser side is responsible for the functions of visualization and interaction, which greatly improves the interpretability with the help of the machine learning side. While the browser side is responsible for the functions of visualization and interaction, which greatly improves the interpretability with the help of the machine learning side.
@ -18,7 +18,7 @@ The code has been placed under the browser side branch https://github.com/505025
## Critical Interface ## Critical Interface
- 1.The researchers need to import target model in the interface StartPro, and run the machine learning side through this interface. - 1.The researchers need to import target model in the interface StartPro, and run the machine learning side through this interface.
- 2.While the program is running, the machine learning side will call the interface oriInputSigni to calculate the interpretable information of the RNN model. - 2.While the program is running, the machine learning side will call the interface oriInputSigni to calculate the interpretable information of DTM.
- 3.The interface getPredict is set to provide the prediction value of the input reordered by the end user. - 3.The interface getPredict is set to provide the prediction value of the input reordered by the end user.
- 4.If you want to use our system to trace the training data, retreating your data is essential, you should use the interface ReadFile to import your model and read your data into cash. The detail of using will show in the branch maching learning side. - 4.If you want to use our system to trace the training data, retreating your data is essential, you should use the interface ReadFile to import your model and read your data into cash. The detail of using will show in the branch maching learning side.
![image](https://github.com/505025234/OrderVis/blob/main/interFace.png) ![image](https://github.com/505025234/OrderVis/blob/main/interFace.png)
@ -30,8 +30,8 @@ The code has been placed under the browser side branch https://github.com/505025
- 3.ReorderByGroup is used to calculate the local interpretability and the glabal interpretability of the model; GetDiffer is used to catch the critical order line with genetic algorithm; GetPartner is used to trace the train data(it cannot be used without pretreatment). - 3.ReorderByGroup is used to calculate the local interpretability and the glabal interpretability of the model; GetDiffer is used to catch the critical order line with genetic algorithm; GetPartner is used to trace the train data(it cannot be used without pretreatment).
- 4.GetPredict is used to calculate the prediction value of the input reordered by the end user. - 4.GetPredict is used to calculate the prediction value of the input reordered by the end user.
### browser side (front-side) ### browser side (front-side)
- 1.GetInput will communicate with the machine learning side and invoke ReorderByGroup, GetDiffer, GetPartner. it will show the interpretable information of the RNN model graphically. - 1.GetInput will communicate with the machine learning side and invoke ReorderByGroup, GetDiffer, GetPartner. it will show the interpretable information of DTM graphically.
- 2.When the end user interact with the system in the form of dragging components or token, DragData will be called and communicate with the machine learning side. DragData will show the prediction value of the reordered input. - 2.When the end user interact with the system in way of dragging components or token, DragData will be called and communicate with the machine learning side. DragData will show the prediction value of the reordered input.
![image](https://github.com/505025234/OrderVis/blob/main/generalizationProcedure.png) ![image](https://github.com/505025234/OrderVis/blob/main/generalizationProcedure.png)
## Example View ## Example View